bug longitude matrices in files for standard latitude/longitude domains when force_coordinate_rotation = True

When attempting to convert some RCM output that has been run over a non rotated domain over Africa with force_coordinate_rotation = True , the real (rather than rotated) longitude matrices have incorrect values (I think out be out by 180 degrees).

I thought that fixing the issues might be as simple as changing
560 rotated_cs = RotatedGeogCS(90., 0., ellipsoid=cs)
to
560 rotated_cs = RotatedGeogCS(90., 180., ellipsoid=cs)

in mip_convert//mip_convert/new_variable.py . Off line testing showed that the correct latitude longitude matrices get produced, however if I try to run get cdds_convert, it crashes with something about an error in cmor.write.

Note I do not have an urgent need for this to be fixed, just reporting in the hope that it can get picked up at some point.

Start in mip_convert/mip_convert/new_variable.py around line 560, where the RotatedGeogCS is created. Reproduce the conversion with non-rotated African RCM output and force_coordinate_rotation = True, then compare the real latitude and longitude matrices. Run cdds_convert and investigate the reported cmor.write failure; done means the matrices are correct and conversion completes successfully.

Written by the indexing model from the issue text.